Medulloblastoma (MB) is the most common malignant brain tumor in children. MB is thought to arise from progenitor cells in the developing cerebellum that fail to exit the cell division cycle properly, thus providing fertile ground for tumor formation. Many genetic anomalies have been identified in human MB, but only a few have been revealed to be causative. We have genetically engineered two different MB-prone mouse strains that lack the cyclin-D-dependent kinase inhibitor, p18lnk4c, a cell cycle regulatory and tumor suppressor protein whose expression was also revealed to be reduced or absent in human MBs. In Specific Aim 1, we propose to document the frequency of inactivation of INK4C/CDKN2C in human MBs and to correlate its loss-of-function with other genetic mutations, gene copy number alterations, and gene expression profiles that define different human MB subsets. We will also explore the combined roles of the C-MYC oncoprotein and p53 tumor suppressor in generating a mouse model of large cell anaplastic MBs, the most aggressive and treatment-resistant form of the disease. Many mouse MBs are characterized by mutations affecting a signaling pathway dominated by the mitogen, Sonic Hedgehog (Shh);genetic alterations affecting the Shh signaling pathway have similarly been documented in a subset of human MBs. Our work has established a role for the bone morphogenic proteins (BMPs) in countering Shh signaling, thereby inhibiting proliferation of mouse MBs, and fostering their neuronal differentiation. Not only do BMPs strongly antagonize MB formation in our mouse models, but downregulation of many BMP-responsive genes is a hallmark of Shh-driven human tumors. In Specific Aim 2, we will study Mathl, a key transcription factor regulated by BMP signaling in an'effort to discern how the activity of this protein is governed, what genes it regulates, and why its role is seemingly essential for MB development. Finally, it is now widely appreciated that small regulatory RNA species (micro-RNAs) globally regulate gene expression and thereby contribute to many forms of cancer. In Specific Aim 3, we propose methods to characterize in detail the micro-RNAs that contribute to MB formation and tumor maintenance in both mice and humans.